20051379 Page 1313 Thursday, November 10, 2005 1:00 PM

COMMENTARY

The many roads to cross-presentation

Tom A.M. Groothuis and Jacques Neefjes

Cross-presentation of extracellular antigens by MHC class I molecules is that usually sampled by the classical required for priming cytotoxic T lymphocytes (CTLs) at locations remote from MHC class I path- the site of infection. Various mechanisms have been proposed to explain cross- way (intracellular antigens within the presentation. One such mechanism involves the fusion of the endoplasmic antigen presenting cell). Hence the reticulum (ER) with the endosomal-phagosomal system, in which the machinery mechanism of antigen degradation and required for peptide loading of MHC class I molecules is introduced directly delivery of the peptide to MHC class I into the . Here, we discuss the evidence for and against the ER- molecules is also likely to be different. phagosome concept as well as other possible mechanisms of cross-presentation. The mechanism of cross-presentation has garnered much interest in recent Downloaded from http://rupress.org/jem/article-pdf/202/10/1313/1153815/jem202101313.pdf by guest on 26 September 2021 The scientific community warmly degradation survives complete destruc- years, in part because cross-presentation receives data that support new theories tion and is transported into the ER by is likely to be important in activating addressing major problems in a field. the peptide transporter TAP (transporter CTLs in response to vaccine antigens. These theories can sometimes become associated with antigen presentation) But the mechanism (or mechanisms) dogma—textbook knowledge—even (2). In the ER, the peptides are loaded has yet to be definitively defined. when based on inconclusive data. Such onto newly synthesized MHC class I data are difficult to correct with essen- molecules, and these complexes are then Deciphering mechanism tially negative results, and such results transported to the cell surface (3). In There are many difficulties inherent in equally are difficult to publish. A recent contrast, the MHC class II processing defining the mechanism of cross-presen- publication by Touret et al. (1) is an ex- pathway is dedicated to the presentation tation. One is the source of antigen. In ception to this rule. In this study, the of exogenous and self-antigens that are vaccine studies, heat shock proteins concept of ER–phagosome fusion was degraded in the endocytic pathway. (such as gp96), apoptotic bodies, the tested and refuted, leaving open several The proteases involved in endocytic content of late endosomes (exosomes), alternative routes for cross-presentation. degradation (cathepsins) are different cell lysates, intact cells, peptides, anti- Here we discuss recent studies on the bi- than those used in the MHC class I bodies, and bead-associated proteins have ology of cross-presentation and explain pathway. Peptides are formed as inter- all been used as sources of antigen (3). why the ER–phagosome mechanism is mediates during late endosomal protein Most of these antigens are extracellular unlikely to contribute to this process. degradation and are loaded onto MHC but are derived from an intracellular class II molecules in a reaction catalyzed source (the infected cell) and are likely Classical and cross-presentation by MHC by the chaperone protein HLA-DM liberated as a result of cell lysis. But in- class I molecules before transport to the plasma membrane tracellular antigens from intact cells can Cross-presentation is the process by (4). MHC class I and MHC class II mol- also be cross-presented. One way this which extracellular antigens, which are ecules thus sample antigenic information could occur is through the swapping of normally presented in association with from different sources, intracellular and intracellular peptides through gap junc- MHC class II molecules, are instead extracellular antigens, respectively. A tions (6). It has been known for a long presented by MHC class I molecules. major exception occurs during cross- time that most tissue cells are electri- This differs from the classical MHC presentation. cally coupled with their neighboring class I processing pathway in which In vivo, DCs—the major cell type cells through gap junctions, small chan- MHC class I molecules present antigens responsible for cross-priming—acquire nels that connect the cytosols of adjacent that are synthesized within the cell. endogenous antigens from infected cells cells. Dendritic cells and activated mono- Classical MHC class I antigen presenta- in the periphery, and then migrate to cytes can also establish gap junctions tion begins with the degradation of the lymph nodes where they display with other cells—including infected intracellularly synthesized proteins by antigenic peptides in association with cells—and thereby acquire antigenic the proteasome. Only a fraction of the MHC class I molecules. MHC class fragments for cross-presentation (6). peptide fragments that result from this I–peptide complexes are recognized by Notably, tumors usually close their gap antigen-specific CTLs, which become junctions, opting to live solitary lives. T.A.M.G. and J.N. are at Div. of Tumor Biology, activated and expand in response to This may explain why tumors often Netherlands Institute, 1066 CX Amsterdam, antigen recognition (5). In this scenario, elicit poor CTL responses. Still, this Netherlands. the source of antigens (intracellular, but immunological coupling through gap CORRESPONDENCE from a different cell, or extracellular as junctions could explain cross-presenta- J.N.: [email protected] in vaccination settings) is distinct from tion under conditions in which the an-

JEM © The Rockefeller University Press $8.00 1313 Vol. 202, No. 10, November 21, 2005 1313–1318 www.jem.org/cgi/doi/10.1084/jem.20051379

20051379 Page 1314 Thursday, November 10, 2005 1:00 PM

tigen-expressing cell does not release available for proteasomal degradation, presentation is not necessarily evidence the antigen into the extracellular milieu. transport into the ER, and presentation for entry of exogenous antigens into the Coupling of antigen-containing cells on MHC class I molecules. Another cytosol of DCs, but it also does not ex- and APCs by gap junctions does not, possibility lies in the observation that clude this possibility. Recent studies re- however, explain how extracellular an- MHC class I molecules can be recycled vealed a role for a putative endocytosis tigens, such as those used in vaccina- from the cell surface along the endocytic signal in MHC class I (12) and for endo- tion studies, are cross-presented. Previ- MHC class II pathway and exchange somal proteases (13) in cross-presenta- ous studies with a variety of extracellular endogenous for exogenous peptides en tion, which support a role for the recy- antigens have demonstrated crucial roles route (10). Notably, TAP and protea- cling pathway. In this model, antigens for TAP (7) and the proteasome (8, 9) in some activities are both required for sur- would be degraded by endocytic pro- cross-presentation. One interpretation face expression of MHC class I mole- teases rather than the proteasome, and of these results is that these antigens (or cules (11), without which the recycling thus some antigens that would normally peptides derived thereof) somehow en- pool cannot exist. Hence, the involve- be presented in the classical pathway ter the cytosol of DCs, making them ment of TAP or proteasomes in cross- might not survive to be cross-presented. Downloaded from http://rupress.org/jem/article-pdf/202/10/1313/1153815/jem202101313.pdf by guest on 26 September 2021

Figure 1. Various models of cross-presentation. In the classical infected cells into the cytosol of DCs. (B) MHC class I molecules can MHC class I antigen–presenting pathway, intracellular antigens are enter the recycling pathway and exchange peptides. (C) ER components degraded by the proteasome and peptidases. A fraction of the resulting become an integral part of the phagosomal pathway. The ERAD pathway peptides associate with TAP in the ER membrane where newly synthe- then exports exogenous antigen from the phagosome into the cytosol sized MHC class I molecules are arrested until loaded with peptide. MHC and phagosomal TAP allows retro-transport of peptides back into the class I–peptide complexes then leave the ER and are transported to the phagosome. (D) Exogenous antigens can be transported over the endoso- plasma membrane. Extracellular antigens can enter this pathway in mal membrane. (E) Exosomes secreted by infected cells can bind to DCs various ways. (A) Gap junctions allow direct transfer of peptides from for cross-presentation.

1314 CROSS-PRESENTATION PATHWAYS | T.A.M. Groothuis and J. Neefjes

20051379 Page 1315 Thursday, November 10, 2005 1:00 PM

COMMENTARY

Other mechanisms besides the recy- Early studies of that live tails utilize the same pathway was not cling pathway might also result in cross- and propagate in suggested addressed. These studies also failed to presentation (Fig. 1). For example, that the phagosomal membrane was satisfactorily address the underlying is- exosomes—vesicles derived from the largely derived from the plasma mem- sue of ER contamination, which ren- interior of endocytic structures that are brane, with a minor contribution from ders the localization of ER-specific released by many cell types—can also other endocytic structures including markers (such as Sec61, TAP, tapasin, induce CTL responses by cross-presen- late endosomes, lysosomes, and MIICs calreticulin, and Erp57) in purified tation (14). Whether simple binding of (vesicles that accumulate MHC class II bead-containing vesicles open to alter- these small vesicles, which contain molecules) (17). But more recent stud- native interpretations. Indeed, Guer- MHC class I–peptide complexes, to the ies—most of which used synthetic monprez et al. used cryo-electron mi- plasma membrane of DCs suffices to beads as a substitute for antigen—found croscopy to detect the ER marker trigger CTL activation is unclear and ER-specific proteins such as calnexin, calreticulin directly by antibodies (10). the mechanism is still poorly defined. calreticulin, and the ERAD translocon They failed to detect these at the In addition, various experiments have subunit Sec61p in the isolated bead- phagosomal membrane but only “in shown that extracellular proteins can be containing fractions (18). Based on this close apposition” in the ER. Downloaded from http://rupress.org/jem/article-pdf/202/10/1313/1153815/jem202101313.pdf by guest on 26 September 2021 transferred from endosomes into the finding, the authors concluded that In the study by Houde et al., the cytosol of DCs (9), although how this the membrane of the bead-containing proteasome and undefined polyubiq- occurs is unclear. It might involve dis- phagosome (beadosome) was derived, uitinated proteins were coisolated with solution of the endocytic membrane or at least in part, from the ER mem- the beadosome, and this resulted in a specific protein transporters that pump brane. But these results could also be rather eccentric model (16). In this the antigen out of endosomes and/or explained by contamination of the be- model, the bead-associated antigens lysosomes. Note that solubilization of adosome membrane with ER-derived are pumped from the beadosome into an antigen-containing endocytic struc- vesicles during purification. the cytosol by the Sec61-containing ture would liberate endosomal proteases Electron microscopy has also been ERAD system (20) and are ubiqui- and likely result in the death of the used to show that the phagosomal tinated by beadosome-associated en- cross-presenting cell. membrane is formed from the ER. zymes during retrotranslocation. In a Gagnon et al. showed that the ER sort of a coupled reaction, the ret- ER–phagosome model membrane and the plasma membrane rotranslocated antigens are degraded by Recently, at least three papers offered fused at the site of bead contact (18). the beadosome-associated proteasome, an alternative model of cross-presenta- However, in that study the content of and the resulting peptide fragments as- tion: direct fusion of phagosomes with the ER lumen did not diffuse into the sociate exclusively with TAP com- the ER membrane. In other words, extracellular medium and membrane- plexes located in the beadosome. If they propose that the phagosomal like structures that separated the ER correct, this suggests that the physical membrane is formed—entirely or in from the phagocytic cup were still visi- laws for Brownian motion do not ap- part—from the ER membrane. As a ble, suggesting the possibility that bona ply to bead-derived antigens since they consequence of this fusion, the enzy- fide fusion did not occur. The authors and their degradation products “know” matic machinery required for the re- also noted that an ER-specific enzy- where to be targeted to: beadosome- lease of phagosomal proteins into the matic activity (glucose-6-phosphatase associated proteasomes and TAP, respec- cytosol (the ER-associated degradation [G6Pase]) was detected in the beado- tively. Ackerman et al. used another [ERAD] system) and the MHC class I some. Since then, several new isotypes approach to test the feasibility of direct loading machinery become an integral of G6Pase have been identified, only fusion of the ER to bead-containing part of the endocytic system (15, 16). one of which contains an ER retention phagosomes (21). They performed the The ERAD system, which shuttles motif (19). Thus, it is possible that this same type of experiments as discussed misfolded proteins from the ER into enzyme might be more widely localized above (with similar problems) but also the cytosol for proteasomal degrada- than it was originally thought to be. showed that a soluble viral TAP inhibi- tion, would thus become the phago- Despite these caveats, the concept tor (US6) could access a macropi- some-to-cytosol protein transporter that the ER contributes to the phago- nocytic compartment and block cross- mentioned earlier. This model offers a somal membrane is highly attractive as presentation of a cointernalized soluble new mechanistic explanation for MHC it provides a mechanistic explanation protein. However, a subsequent paper class I cross-presentation, but has re- for the cross-presentation of extracellu- by the same authors showed that exog- cently been tested and refuted (1). lar proteins. More recently, two papers enous proteins could follow a retro- Moreover, we suggest that this model were published claiming that this route grade transport pathway from endo- is problematic for other reasons and is was operational when antigens were somes, through the Golgi and back thus unlikely to contribute significantly given in association with 3-m beads into the ER (22). This suggests that to cross-presentation in vivo. (15, 16). Whether other antigen cock- soluble proteins might be cross-pre-

JEM VOL. 202, November 21, 2005 1315

20051379 Page 1316 Thursday, November 10, 2005 1:00 PM

sented as a result of their ability to di- significant contribution by the ER (es- quired for stimulation of primed and rectly access the MHC class I process- timated between 0 and 10%) (1). naive T cells, respectively (28). Based ing machinery in the ER lumen. on our calculations, T cell activation Retrograde transport through the Calculating the odds of ER-phagosome would thus occur only if nearly all the Golgi could thus explain how soluble fusion peptides delivered into the cytosol antibodies are able to gain access to the The feasibility of this model may be from bead-derived phagosomes found ER and why soluble US6 inhibits deduced by “number crunching” ac- their way back into phagosomes with- cross-presentation. In other words, it cording to Yewdellian philosophy (25). out any competition from endogenous does not prove the existence of a An estimation of the numbers of MHC peptides. Given that endogenous pep- mixed ER–phagosome compartment, class I molecules entering a 3-m bead- tides are present in the cytoplasm even but rather reveals yet another potential osome (15, 16) from the ER and the before exogenous antigens are de- mechanism by which extracellular anti- numbers of molecules required to ini- graded (2), beadosome-derived anti- gens could be cross-presented. tiate cross-presentation may help reveal genic peptides would likely be out- So does the ER–phagosome exist? whether the ER–phagosome fusion competed by endogenous peptides and As mentioned earlier, experiments us- mechanism is plausible. If ER-derived would thus return to the beadosomes Downloaded from http://rupress.org/jem/article-pdf/202/10/1313/1153815/jem202101313.pdf by guest on 26 September 2021 ing bacterial phagosomes suggested that membrane constituted 10% of the phago- too late to load the few ER-derived the plasma membrane rather than the somal membrane, would this suffice MHC class I molecules at that location. ER was the primary source of phagoso- for efficient cross-presentation? If one However, if the ER translocon protein mal membranes (17, 23, 24). In their approximates that an average rounded Sec61, which also inserts polypeptides new study, Touret et al. rigorously cell has a radius of 10 m, then by into the ER lumen as they are trans- tested the origin of the phagosomal comparing the surface area (4r2) of lated, is introduced in the phagosome, membrane around ingested beads or the cell with that of the 3-m bead- as suggested by previous studies (15, bacteria (1). The experiments per- induced phagosome, one can calculate 16, 18), then translation of novel pro- formed in earlier studies (15, 16, 18) that 2% ([1.5/10]2 or 2 102) of teins could continue at the phagosomal were largely repeated by Touret et al. the plasma membrane is contributed to membrane and result in the deposition Quantitative immunolabeling for ER the membrane of the 3-m beadosome of de novo–translated proteins directly markers and G6Pase activity measure- (radius of 1.5 m). The ER consti- into the phagosome. The phagosome ments showed no evidence for a contri- tutes 60% of cellular lipids (compared would thus mimic the ER by allowing bution of the ER to the phagosomal with the plasma membrane’s 5%) (26), the introduction of nascent proteins membrane. In addition, extracellular which means that 0.16% (2 · 102 (including MHC class I molecules), but dyes were shown to be constrained in [5/60]) of the ER membrane would be no data are available on this point. the phagosome and did not diffuse into donated to each beadosome, if the be- Our own electron microscopy the ER, as would be expected if the adosome membrane was composed en- analysis shows that typical ribosomal ER–lumen was in (even temporal) con- tirely of ER-derived membrane. structures can be found associated to tinuum with the forming beadosome. MHC class I molecules have a half- the ER membrane, but no such struc- The authors performed a plethora of life of over 12 h (although this varies tures were localized to the beadoso- experiments, none of which provided somewhat in different cell types) but mal membrane (Fig. 2). Bead-induced evidence for ER–phagosome fusion (1). are available for peptide loading in the cross-presentation therefore must rely One experiment deserves special at- ER for 30 min or less (27). Even with on the few ER-derived MHC class I tention because of its elegance. In this this conservative estimation, this im- molecules, or on cell surface–derived experiment, the biotin-binding pro- plies that at every moment less than MHC class I molecules (2 million 2 · tein avidin was expressed with an ER 100,000 peptide-receptive MHC class I 102 40,000 surface MHC class I retention signal (KDEL) in a macro- molecules would be located in the ER molecules in a 3-m beadosome) for phage cell line. Beads coated with bi- of a DC that contains a total of 2 mil- cross-presentation of phagosomal anti- otin were then phagocytosed by the lion MHC class I molecules. Of these, gens. Surface MHC class I molecules cells, and direct contact between the 16 (100,000 2 · 102 [5/60] can efficiently exchange peptides be- ER-retained avidin and bead-associ- 0.1) MHC class I molecules would en- tween pH 4.5 and 5.5 (10), suggesting ated biotin was quantified. No avidin– ter one 3-m phagosome at a 10% that peptide loading could take place in biotin interactions were observed, cast- contribution of the ER and 160 mole- the phagosomal environment, although ing more doubt on the concept of cules at a 100% contribution. These the loading would be considerably less direct interaction between the ER and numbers would decrease ninefold for a efficient than in the “specialized” phagosomes (1). Touret et al. conclude 1-m bead and even more for soluble MHC class I–loading complex in the that the plasma membrane and the en- antigens and immune complexes. ER. However, inefficiency would not docytic pathway are the major sources A minimum of 40 and 400 MHC– be a major issue in the recycling path- for phagosomal membranes, with no peptide complexes is reported to be re- way as 40,000 surface-derived MHC

1316 CROSS-PRESENTATION PATHWAYS | T.A.M. Groothuis and J. Neefjes

20051379 Page 1317 Thursday, November 10, 2005 1:00 PM

COMMENTARY

tion of the antigen-processing machinery. Immunol. Rev. 207:60–76. 4. Bryant, P.W., A.M. Lennon-Dumenil, E. Fiebiger, C. Lagaudriere-Gesbert, and H.L. Ploegh. 2002. Proteolysis and antigen pre- sentation by MHC class II molecules. Adv. Immunol. 80:71–114. 5. Heath, W.R., G.T. Belz, G.M. Behrens, C.M. Smith, S.P. Forehan, I.A. Parish, G.M. Davey, N.S. Wilson, F.R. Carbone, and J.A. Villadangos. 2004. Cross-presenta- tion, dendritic cell subsets, and the genera- tion of immunity to cellular antigens. Immu- nol. Rev. 199:9–26. 6. Neijssen, J., C. Herberts, J.W. Drijfhout, E. Reits, L. Janssen, and J. Neefjes. 2005. Cross-presentation by intercellular peptide Downloaded from http://rupress.org/jem/article-pdf/202/10/1313/1153815/jem202101313.pdf by guest on 26 September 2021 Figure 2. Beadosomes, ER, ribosomes, and mitochondria in a DC. 1-m latex beads were transfer through gap junctions. Nature. 434: 83–88. endocytosed by human monocyte-derived DCs for 30 min before processing for electron microscopy. 7. Huang, A.Y., A.T. Bruce, D.M. Pardoll, The cells were fixed with a mixture of formaldehyde and glutaraldehyde and embedded in epon for and H.I. Levitsky. 1996. In vivo cross-prim- thin sectioning. The bead, mitochondria (Mt), and two ER profiles are indicated. The arrowheads point ing of MHC class I-restricted antigens to ribosomes in association to the ER membrane. No such structures are observed on the beadosomal requires the TAP transporter. Immunity. membrane. The box denotes a position in the section with lower resolution where the ER is apposite 4:349–355. to the mitochondrion. Bar, 200 nm. 8. Ruedl, C., T. Storni, F. Lechner, T. Bachi, and M.F. Bachmann. 2002. Cross-presenta- tion of virus-like particles by skin-derived class I molecules would be available to biological question, cross-presentation CD8() dendritic cells: a dispensable role bind peptides, whereas only 160 MHC has direct consequences for vaccination for TAP. Eur. J. Immunol. 32:818–825. 9. Norbury, C.C., L.J. Hewlett, A.R. Prescott, class I molecules would be available if strategies aimed at inducing CTL re- N. Shastri, and C. Watts. 1995. Class I the beadosome membrane were de- sponses. These vaccines should be able MHC presentation of exogenous soluble an- rived entirely from the ER. to induce potent CTL responses and T tigen via macropinocytosis in bone marrow cell memory, and the specificity of the macrophages. Immunity. 3:783–791. CTL response will result from the 10. Gromme, M., F.G. Uytdehaag, H. Janssen, Concluding considerations J. Calafat, R.S. van Binnendijk, M.J. Various studies have reached diametri- cross-presentation of antigenic fragments. Kenter, A. Tulp, D. Verwoerd, and J. cally opposing conclusions regarding Understanding the mechanism(s) of Neefjes. 1999. Recycling MHC class I mol- ecules and endosomal peptide loading. Proc. the origin of the phagosomal mem- cross-presentation will help rationalize vaccine development and improve the Natl. Acad. Sci. USA. 96:10326–10331. brane—a crucial question underlying 11. Chefalo, P.J., A.G. Grandea III, L. Van the mechanism of cross-presentation. chances to arrive at successful antiviral Kaer, and C.V. Harding. 2003. Tapasin/ Gagnon et al. showed an ER contribu- and antitumor vaccines. and TAP1/ macrophages are deficient in vacuolar alternate class I MHC (MHC-I) tion to the phagosomal membrane in This work was supported by grants from the Dutch processing due to decreased MHC-I stability macrophages, but not in other cells Cancer Society KWF. We thank Alexander Griekspoor, at phagolysosomal pH. J. Immunol. 170: such as neutrophils (18). These experi- Helen Pickersgill, and Andra Neefjes-Borst for critically 5825–5833. ments could also not be confirmed in reading the manuscript. We thank H. Janssen for the 12. Lizee, G., G. Basha, J. Tiong, J.P. Julien, M. electron microscopy. macrophages and DCs in the recent Tian, K.E. Biron, and W.A. Jefferies. 2003. Control of dendritic cell cross-presentation study by Touret et al. (1). Although the REFERENCES by the major histocompatibility complex studies that argue for the ER–phagoso- 1. Touret, N., P. Paroutis, M. Terbiznik, R. class I cytoplasmic domain. Nat. Immunol. mal fusion mechanism received much Harrison, S. Trombetta, M. Pypaert, A. 4:1065–1073. attention, they did not address the Chow, A. Jiang, J. Shaw, C. Yip, et al. 13. Shen, L., L.J. Sigal, M. Boes, and K.L. 2005. Quantitative and dynamic assessment Rock. 2004. Important role of cathepsin S mechanism of cross-presentation of of the contribution of the endoplasmic retic- in generating peptides for TAP-independent physiological antigens, including anti- ulum to phagosome formation. Cell. 123: MHC class I crosspresentation in vivo. Im- body-bound antigens (29), soluble an- 157–170. munity. 21:155–165. tigens (30), and intracellular antigens 2. Reits, E., A. Griekspoor, J. Neijssen, T. 14. Wolfers, J., A. Lozier, G. Raposo, A. Reg- (6), which might all follow distinct Groothuis, K. Jalink, P. van Veelen, H. Jan- nault, C. Thery, C. Masurier, C. Flament, ssen, J. Calafat, J.W. Drijfhout, and J. S. Pouzieux, F. Faure, T. Tursz, et al. 2001. pathways of degradation inside the cell. Neefjes. 2003. Peptide diffusion, protection, Tumor-derived exosomes are a source of Furthermore, a simple calculation pre- and degradation in nuclear and cytoplasmic shared tumor rejection antigens for CTL dicts that cross-presentation via the compartments before antigen presentation cross-priming. Nat. Med. 7:297–303. ER–phagosome pathway would be by MHC class I. Immunity. 18:97–108. 15. Guermonprez, P., L. Saveanu, M. Kleijmeer, 3. Groothuis, T.A., A.C. Griekspoor, J.J. J. Davoust, P. Van Endert, and S. Amig- highly inefficient, if at all possible. Neijssen, C.A. Herberts, and J.J. Neefjes. orena. 2003. ER-phagosome fusion defines Apart from being an interesting 2005. MHC class I alleles and their explora- an MHC class I cross-presentation compart-

JEM VOL. 202, November 21, 2005 1317

20051379 Page 1318 Thursday, November 10, 2005 1:00 PM

ment in dendritic cells. Nature. 425:397–402. ated transfer of a membrane protein from Roberts, and P. Walter. 2002. Molecular Biol- 16. Houde, M., S. Bertholet, E. Gagnon, S. the endoplasmic reticulum to the protea- ogy of the Cell. 4th edition. Garland Science, Brunet, G. Goyette, A. Laplante, M.F. Prin- some for destruction. Nature. 384:432–438. New York. 661 pp. ciotta, P. Thibault, D. Sacks, and M. Desjar- 21. Ackerman, A.L., C. Kyritsis, R. Tampe, and 27. Neefjes, J.J., G.J. Hammerling, and F. Mom- dins. 2003. Phagosomes are competent P. Cresswell. 2003. Early phagosomes in burg. 1993. Folding and assembly of major organelles for antigen cross-presentation. dendritic cells form a cellular compartment histocompatibility complex class I het- Nature. 425:402–406. sufficient for cross presentation of exoge- erodimers in the endoplasmic reticulum of 17. Touret, N., P. Paroutis, and S. Grinstein. nous antigens. Proc. Natl. Acad. Sci. USA. intact cells precedes the binding of peptide. 2005. The nature of the phagosomal mem- 100:12889–12894. J. Exp. Med. 178:1971–1980. brane: endoplasmic reticulum versus plasma- 22. Ackerman, A.L., C. Kyritsis, R. Tampe, and 28. Kimachi, K., M. Croft, and H.M. Grey. lemma. J. Leukoc. Biol. 77:878–885. P. Cresswell. 2005. Access of soluble anti- 1997. The minimal number of antigen-major 18. Gagnon, E., S. Duclos, C. Rondeau, E. gens to the endoplasmic reticulum can ex- histocompatibility complex class II complexes Chevet, P.H. Cameron, O. Steele-Mor- plain cross-presentation by dendritic cells. required for activation of naive and primed T timer, J. Paiement, J.J. Bergeron, and M. Nat. Immunol. 6:107–113. cells. Eur. J. Immunol. 27:3310–3317. Desjardins. 2002. Endoplasmic reticulum- 23. Russell, D.G. 2003. Phagosomes, fatty acids 29. den Haan, J.M., and M.J. Bevan. 2002. mediated phagocytosis is a mechanism of and tuberculosis. Nat. Cell Biol. 5:776–778. Constitutive versus activation-dependent entry into macrophages. Cell. 110:119–131. 24. Cossart, P., and P.J. Sansonetti. 2004. Bacte- cross-presentation of immune complexes by 19. Guionie, O., E. Clottes, K. Stafford, and A. rial invasion: the paradigms of enteroinva- CD8() and CD8() dendritic cells in Downloaded from http://rupress.org/jem/article-pdf/202/10/1313/1153815/jem202101313.pdf by guest on 26 September 2021 Burchell. 2003. Identification and character- sive . Science. 304:242–248. vivo. J. Exp. Med. 196:817–827. isation of a new human glucose-6-phos- 25. Yewdell, J.W. 2001. Not such a dismal sci- 30. Kurts, C., H. Kosaka, F.R. Carbone, J.F. phatase isoform. FEBS Lett. 551:159–164. ence: the economics of protein synthesis, Miller, and W.R. Heath. 1997. Class I-restricted 20. Wiertz, E.J., D. Tortorella, M. Bogyo, J. folding, degradation and antigen processing. cross-presentation of exogenous self-anti- Yu, W. Mothes, T.R. Jones, T.A. Rapo- Trends Cell Biol. 11:294–297. gens leads to deletion of autoreactive CD8() port, and H.L. Ploegh. 1996. Sec61-medi- 26. Alberts, B., A. Johnson, J. Lewis, M. Raff, K. T cells. J. Exp. Med. 186:239–245.

1318 CROSS-PRESENTATION PATHWAYS | T.A.M. Groothuis and J. Neefjes